As for making eras last longer in general, there are a few things you can do about that. In my opinion the problem isn't that research is too fast in general, it's that the later techs are just too cheap relative to the ever-increasing science output of your empire. The march patch helped with this, but it can go further.
For instance, consider the research buildings:
(base: 1 science per population)
Library: +1 science per 2 population
University: +50% science
Observatory: +50% science, must have a mountain nearby
Public School: +1 science per population
Research Lab: +100% science
So by the time you get to the end of the game, you'll be producing (1 + 0.5 + 1)*(1 + 0.5 + 0.5 + 1.0) = 7.5 science per population for cities that have observatories. Without that it's 6.25, but any other +50% (like the National College) gets you back to 7.5 again.
There are a few other sources of beakers: the Academy, Scientist specialists, trading posts with the right policy. And most of the Rationalism tree is about boosting science. But the dominant source is still the amount I listed above.
So what can you do about it? In my mod, I did this:
Library: +1 science per THREE population, and +1 culture
University: +30% science, +1 culture
Observatory: +20% science, +3 culture, still needs a mountain
Public School: +40% science, +1 culture
Research Lab: +50% science, +10% gold
Medical Lab: +10% science, and the food storage bit was dropped to 20%.
Finally, the Secularism policy was reduced from "+2 science per specialist" to +1.
Now, instead of 7.5 beakers per population, you peak at 3.33. As a result, your research speed in the late game is halved, with earlier eras seeing a bit less of a proportional decrease.
There's a secondary problem, where some buildings and units are just too expensive. I solved that by tweaking military buildings and such. For instance, in my own mod, instead of +15 XP for land units, the Armory now gives +10% to unit production and +10 XP for land units. Similar change for the Military Academy. But the underlying research rates were the main issue.